Apple Inc.

University of Washington

Saman Naderiparizi

Sayyed Karen Khatamifard

Mohammad Rastegari

Joshua R Smith

Shyamnath Gollakota

Vamsi Talla

Rohit Kumar Gupta

Ramana V Rachakonda

Hessam Bagherinezhad

David S Bakin

Anish Jnyaneshwar Prabhu

Bryce Kellogg

Sophie Lebrecht

Andrew Luke Chronister

Alon Yaakov

Mehrdad Hessar

Jeffrey D Marker

Chenfan Sun

Ben Ransford

Husam Khashiboun

James C Gabriel

Jonathan Gelsey

Andrew Z Luo

In one embodiment, a computing device includes an input sensor providing an input data; a programmable logic device (PLD) implementing a convolutional neural network (CNN), wherein: each compute block of the PLD corresponds to one of a multiple of convolutional layers of the CNN, each compute block of the PLD is placed in proximity to at least two memory blocks, a first one of the memory blocks serves as a buffer for the corresponding layer of the CNN, and a second one of the memory blocks stores model-specific parameters for the corresponding layer of the CNN.

Customizable chip for AI applications

Embodiments relate to streaming convolution operations in a neural processor circuit that includes a neural engine circuit and a neural task manager. The neural task manager obtains multiple task descriptors and multiple subtask descriptors. Each task descriptor identifies a respective set of the convolution operations of a respective layer of a set of layers. Each subtask descriptor identifies a corresponding task descriptor and a subset of the convolution operations on a portion of a layer of the set of layers identified by the corresponding task descriptor. The neural processor circuit configures the neural engine circuit for execution of the subset of the convolution operations using the corresponding task descriptor. The neural engine circuit performs the subset of the convolution operations to generate output data that correspond to input data of another subset of the convolution operations identified by another subtask descriptor from the list of subtask descriptors.

Subtask storage for streaming convolutions in neural network processor

Systems and processes for training and compressing a convolutional neural network model include the use of quantization and layer fusion. Quantized training data is passed through a convolutional layer of a neural network model to generate convolutional results during a first iteration of training the neural network model. The convolutional results are passed through a batch normalization layer of the neural network model to update normalization parameters of the batch normalization layer. The convolutional layer is fused with the batch normalization layer to generate a first fused layer and the fused parameters of the fused layer are quantized. The quantized training data is passed through the fused layer using the quantized fused parameters to generate output data, which may be quantized for a subsequent layer in the training iteration.

Compressed convolutional neural network models

Examples described herein include systems, devices, and methods for backscattering carrier signals in accordance with pixel values of an image and/or video. Signals having a property proportionate to pixel values may be converted into a pulse-containing waveform having pulses whose widths and/or duty cycles are determined based on the pixel values. Backscatter transmitters may backscatter a carrier signal in accordance with the pulse-containing waveform to provide the pixel values to a receiver. In this manner, video transmission at low power and/or battery-free operation may be provided.

Image and/or video transmission using backscatter devices

Examples described herein include wireless transmitters configured for power transmission. Example wireless transmitters may insert power packets into wireless communications such that power harvesting circuitry may harvest sufficiently continuous power from the wireless communication signals. Example power harvesting circuitry is configured to harvest power across multiple wireless communication channels. Example chargers are described which may harvest power from wireless communication signals (e.g. Wi-Fi signals).

Power transmission using wireless communication signals

Growing community of inventors

Seattle, WA, United States of America

Saman Naderiparizi